[deliverable/linux.git] / net / ipv4 / inet_fragment.c

/*
 * inet fragments management
 *
 *		This program is free software; you can redistribute it and/or
 *		modify it under the terms of the GNU General Public License
 *		as published by the Free Software Foundation; either version
 *		2 of the License, or (at your option) any later version.
 *
 * 		Authors:	Pavel Emelyanov <xemul@openvz.org>
 *				Started as consolidation of ipv4/ip_fragment.c,
 *				ipv6/reassembly. and ipv6 nf conntrack reassembly
 */

#include <linux/list.h>
#include <linux/spinlock.h>
#include <linux/module.h>
#include <linux/timer.h>
#include <linux/mm.h>
#include <linux/random.h>
#include <linux/skbuff.h>
#include <linux/rtnetlink.h>
#include <linux/slab.h>

#include <net/sock.h>
#include <net/inet_frag.h>

static void inet_frag_secret_rebuild(unsigned long dummy)
{
	struct inet_frags *f = (struct inet_frags *)dummy;
	unsigned long now = jiffies;
	int i;

	write_lock(&f->lock);
	get_random_bytes(&f->rnd, sizeof(u32));
	for (i = 0; i < INETFRAGS_HASHSZ; i++) {
		struct inet_frag_queue *q;
		struct hlist_node *n;

		hlist_for_each_entry_safe(q, n, &f->hash[i], list) {
			unsigned int hval = f->hashfn(q);

			if (hval != i) {
				hlist_del(&q->list);

				/* Relink to new hash chain. */
				hlist_add_head(&q->list, &f->hash[hval]);
			}
		}
	}
	write_unlock(&f->lock);

	mod_timer(&f->secret_timer, now + f->secret_interval);
}

void inet_frags_init(struct inet_frags *f)
{
	int i;

	for (i = 0; i < INETFRAGS_HASHSZ; i++)
		INIT_HLIST_HEAD(&f->hash[i]);

	rwlock_init(&f->lock);

	f->rnd = (u32) ((num_physpages ^ (num_physpages>>7)) ^
				   (jiffies ^ (jiffies >> 6)));

	setup_timer(&f->secret_timer, inet_frag_secret_rebuild,
			(unsigned long)f);
	f->secret_timer.expires = jiffies + f->secret_interval;
	add_timer(&f->secret_timer);
}
EXPORT_SYMBOL(inet_frags_init);

void inet_frags_init_net(struct netns_frags *nf)
{
	nf->nqueues = 0;
	init_frag_mem_limit(nf);
	INIT_LIST_HEAD(&nf->lru_list);
	spin_lock_init(&nf->lru_lock);
}
EXPORT_SYMBOL(inet_frags_init_net);

void inet_frags_fini(struct inet_frags *f)
{
	del_timer(&f->secret_timer);
}
EXPORT_SYMBOL(inet_frags_fini);

void inet_frags_exit_net(struct netns_frags *nf, struct inet_frags *f)
{
	nf->low_thresh = 0;

	local_bh_disable();
	inet_frag_evictor(nf, f, true);
	local_bh_enable();

	percpu_counter_destroy(&nf->mem);
}
EXPORT_SYMBOL(inet_frags_exit_net);

static inline void fq_unlink(struct inet_frag_queue *fq, struct inet_frags *f)
{
	write_lock(&f->lock);
	hlist_del(&fq->list);
	fq->net->nqueues--;
	write_unlock(&f->lock);
	inet_frag_lru_del(fq);
}

void inet_frag_kill(struct inet_frag_queue *fq, struct inet_frags *f)
{
	if (del_timer(&fq->timer))
		atomic_dec(&fq->refcnt);

	if (!(fq->last_in & INET_FRAG_COMPLETE)) {
		fq_unlink(fq, f);
		atomic_dec(&fq->refcnt);
		fq->last_in |= INET_FRAG_COMPLETE;
	}
}
EXPORT_SYMBOL(inet_frag_kill);

static inline void frag_kfree_skb(struct netns_frags *nf, struct inet_frags *f,
		struct sk_buff *skb)
{
	if (f->skb_free)
		f->skb_free(skb);
	kfree_skb(skb);
}

void inet_frag_destroy(struct inet_frag_queue *q, struct inet_frags *f,
					int *work)
{
	struct sk_buff *fp;
	struct netns_frags *nf;
	unsigned int sum, sum_truesize = 0;

	WARN_ON(!(q->last_in & INET_FRAG_COMPLETE));
	WARN_ON(del_timer(&q->timer) != 0);

	/* Release all fragment data. */
	fp = q->fragments;
	nf = q->net;
	while (fp) {
		struct sk_buff *xp = fp->next;

		sum_truesize += fp->truesize;
		frag_kfree_skb(nf, f, fp);
		fp = xp;
	}
	sum = sum_truesize + f->qsize;
	if (work)
		*work -= sum;
	sub_frag_mem_limit(q, sum);

	if (f->destructor)
		f->destructor(q);
	kfree(q);

}
EXPORT_SYMBOL(inet_frag_destroy);

int inet_frag_evictor(struct netns_frags *nf, struct inet_frags *f, bool force)
{
	struct inet_frag_queue *q;
	int work, evicted = 0;

	if (!force) {
		if (frag_mem_limit(nf) <= nf->high_thresh)
			return 0;
	}

	work = frag_mem_limit(nf) - nf->low_thresh;
	while (work > 0) {
		spin_lock(&nf->lru_lock);

		if (list_empty(&nf->lru_list)) {
			spin_unlock(&nf->lru_lock);
			break;
		}

		q = list_first_entry(&nf->lru_list,
				struct inet_frag_queue, lru_list);
		atomic_inc(&q->refcnt);
		spin_unlock(&nf->lru_lock);

		spin_lock(&q->lock);
		if (!(q->last_in & INET_FRAG_COMPLETE))
			inet_frag_kill(q, f);
		spin_unlock(&q->lock);

		if (atomic_dec_and_test(&q->refcnt))
			inet_frag_destroy(q, f, &work);
		evicted++;
	}

	return evicted;
}
EXPORT_SYMBOL(inet_frag_evictor);

static struct inet_frag_queue *inet_frag_intern(struct netns_frags *nf,
		struct inet_frag_queue *qp_in, struct inet_frags *f,
		void *arg)
{
	struct inet_frag_queue *qp;
#ifdef CONFIG_SMP
#endif
	unsigned int hash;

	write_lock(&f->lock);
	/*
	 * While we stayed w/o the lock other CPU could update
	 * the rnd seed, so we need to re-calculate the hash
	 * chain. Fortunatelly the qp_in can be used to get one.
	 */
	hash = f->hashfn(qp_in);
#ifdef CONFIG_SMP
	/* With SMP race we have to recheck hash table, because
	 * such entry could be created on other cpu, while we
	 * promoted read lock to write lock.
	 */
	hlist_for_each_entry(qp, &f->hash[hash], list) {
		if (qp->net == nf && f->match(qp, arg)) {
			atomic_inc(&qp->refcnt);
			write_unlock(&f->lock);
			qp_in->last_in |= INET_FRAG_COMPLETE;
			inet_frag_put(qp_in, f);
			return qp;
		}
	}
#endif
	qp = qp_in;
	if (!mod_timer(&qp->timer, jiffies + nf->timeout))
		atomic_inc(&qp->refcnt);

	atomic_inc(&qp->refcnt);
	hlist_add_head(&qp->list, &f->hash[hash]);
	nf->nqueues++;
	write_unlock(&f->lock);
	inet_frag_lru_add(nf, qp);
	return qp;
}

static struct inet_frag_queue *inet_frag_alloc(struct netns_frags *nf,
		struct inet_frags *f, void *arg)
{
	struct inet_frag_queue *q;

	q = kzalloc(f->qsize, GFP_ATOMIC);
	if (q == NULL)
		return NULL;

	q->net = nf;
	f->constructor(q, arg);
	add_frag_mem_limit(q, f->qsize);

	setup_timer(&q->timer, f->frag_expire, (unsigned long)q);
	spin_lock_init(&q->lock);
	atomic_set(&q->refcnt, 1);

	return q;
}

static struct inet_frag_queue *inet_frag_create(struct netns_frags *nf,
		struct inet_frags *f, void *arg)
{
	struct inet_frag_queue *q;

	q = inet_frag_alloc(nf, f, arg);
	if (q == NULL)
		return NULL;

	return inet_frag_intern(nf, q, f, arg);
}

struct inet_frag_queue *inet_frag_find(struct netns_frags *nf,
		struct inet_frags *f, void *key, unsigned int hash)
	__releases(&f->lock)
{
	struct inet_frag_queue *q;
	int depth = 0;

	hlist_for_each_entry(q, &f->hash[hash], list) {
		if (q->net == nf && f->match(q, key)) {
			atomic_inc(&q->refcnt);
			read_unlock(&f->lock);
			return q;
		}
		depth++;
	}
	read_unlock(&f->lock);

	if (depth <= INETFRAGS_MAXDEPTH)
		return inet_frag_create(nf, f, key);
	else
		return ERR_PTR(-ENOBUFS);
}
EXPORT_SYMBOL(inet_frag_find);

void inet_frag_maybe_warn_overflow(struct inet_frag_queue *q,
				   const char *prefix)
{
	static const char msg[] = "inet_frag_find: Fragment hash bucket"
		" list length grew over limit " __stringify(INETFRAGS_MAXDEPTH)
		". Dropping fragment.\n";

	if (PTR_ERR(q) == -ENOBUFS)
		LIMIT_NETDEBUG(KERN_WARNING "%s%s", prefix, msg);
}
EXPORT_SYMBOL(inet_frag_maybe_warn_overflow);
Commit	Line	Data
7eb95156 PE	1	/*
	2	* inet fragments management
	3	*
	4	* This program is free software; you can redistribute it and/or
	5	* modify it under the terms of the GNU General Public License
	6	* as published by the Free Software Foundation; either version
	7	* 2 of the License, or (at your option) any later version.
	8	*
	9	* Authors: Pavel Emelyanov <xemul@openvz.org>
	10	* Started as consolidation of ipv4/ip_fragment.c,
	11	* ipv6/reassembly. and ipv6 nf conntrack reassembly
	12	*/
	13
	14	#include <linux/list.h>
	15	#include <linux/spinlock.h>
	16	#include <linux/module.h>
	17	#include <linux/timer.h>
	18	#include <linux/mm.h>
321a3a99	19	#include <linux/random.h>
1e4b8287 PE	20	#include <linux/skbuff.h>
1e4b8287 PE	21	#include <linux/rtnetlink.h>
5a0e3ad6	22	#include <linux/slab.h>
7eb95156	23
5a3da1fe	24	#include <net/sock.h>
7eb95156 PE	25	#include <net/inet_frag.h>
7eb95156 PE	26
321a3a99 PE	27	static void inet_frag_secret_rebuild(unsigned long dummy)
	28	{
	29	struct inet_frags f = (struct inet_frags )dummy;
	30	unsigned long now = jiffies;
	31	int i;
	32
	33	write_lock(&f->lock);
	34	get_random_bytes(&f->rnd, sizeof(u32));
	35	for (i = 0; i < INETFRAGS_HASHSZ; i++) {
	36	struct inet_frag_queue *q;
b67bfe0d	37	struct hlist_node *n;
321a3a99	38
b67bfe0d	39	hlist_for_each_entry_safe(q, n, &f->hash[i], list) {
321a3a99 PE	40	unsigned int hval = f->hashfn(q);
	41
	42	if (hval != i) {
	43	hlist_del(&q->list);
	44
	45	/* Relink to new hash chain. */
	46	hlist_add_head(&q->list, &f->hash[hval]);
	47	}
	48	}
	49	}
	50	write_unlock(&f->lock);
	51
3b4bc4a2	52	mod_timer(&f->secret_timer, now + f->secret_interval);
321a3a99 PE	53	}
321a3a99 PE	54
7eb95156 PE	55	void inet_frags_init(struct inet_frags *f)
	56	{
	57	int i;
	58
	59	for (i = 0; i < INETFRAGS_HASHSZ; i++)
	60	INIT_HLIST_HEAD(&f->hash[i]);
	61
7eb95156 PE	62	rwlock_init(&f->lock);
	63
	64	f->rnd = (u32) ((num_physpages ^ (num_physpages>>7)) ^
	65	(jiffies ^ (jiffies >> 6)));
	66
b24b8a24 PE	67	setup_timer(&f->secret_timer, inet_frag_secret_rebuild,
b24b8a24 PE	68	(unsigned long)f);
3b4bc4a2	69	f->secret_timer.expires = jiffies + f->secret_interval;
321a3a99	70	add_timer(&f->secret_timer);
7eb95156 PE	71	}
	72	EXPORT_SYMBOL(inet_frags_init);
	73
e5a2bb84 PE	74	void inet_frags_init_net(struct netns_frags *nf)
	75	{
	76	nf->nqueues = 0;
d433673e	77	init_frag_mem_limit(nf);
3140c25c	78	INIT_LIST_HEAD(&nf->lru_list);
3ef0eb0d	79	spin_lock_init(&nf->lru_lock);
e5a2bb84 PE	80	}
	81	EXPORT_SYMBOL(inet_frags_init_net);
	82
7eb95156 PE	83	void inet_frags_fini(struct inet_frags *f)
7eb95156 PE	84	{
321a3a99	85	del_timer(&f->secret_timer);
7eb95156 PE	86	}
7eb95156 PE	87	EXPORT_SYMBOL(inet_frags_fini);
277e650d	88
81566e83 PE	89	void inet_frags_exit_net(struct netns_frags nf, struct inet_frags f)
	90	{
	91	nf->low_thresh = 0;
e8e16b70 DM	92
e8e16b70 DM	93	local_bh_disable();
6b102865	94	inet_frag_evictor(nf, f, true);
e8e16b70	95	local_bh_enable();
6d7b857d JDB	96
6d7b857d JDB	97	percpu_counter_destroy(&nf->mem);
81566e83 PE	98	}
	99	EXPORT_SYMBOL(inet_frags_exit_net);
	100
277e650d PE	101	static inline void fq_unlink(struct inet_frag_queue fq, struct inet_frags f)
	102	{
	103	write_lock(&f->lock);
	104	hlist_del(&fq->list);
e5a2bb84	105	fq->net->nqueues--;
277e650d	106	write_unlock(&f->lock);
3ef0eb0d	107	inet_frag_lru_del(fq);
277e650d PE	108	}
	109
	110	void inet_frag_kill(struct inet_frag_queue fq, struct inet_frags f)
	111	{
	112	if (del_timer(&fq->timer))
	113	atomic_dec(&fq->refcnt);
	114
bc578a54	115	if (!(fq->last_in & INET_FRAG_COMPLETE)) {
277e650d PE	116	fq_unlink(fq, f);
277e650d PE	117	atomic_dec(&fq->refcnt);
bc578a54	118	fq->last_in \|= INET_FRAG_COMPLETE;
277e650d PE	119	}
277e650d PE	120	}
277e650d	121	EXPORT_SYMBOL(inet_frag_kill);
1e4b8287	122
6ddc0822	123	static inline void frag_kfree_skb(struct netns_frags nf, struct inet_frags f,
d433673e	124	struct sk_buff *skb)
1e4b8287	125	{
1e4b8287 PE	126	if (f->skb_free)
	127	f->skb_free(skb);
	128	kfree_skb(skb);
	129	}
	130
	131	void inet_frag_destroy(struct inet_frag_queue q, struct inet_frags f,
	132	int *work)
	133	{
	134	struct sk_buff *fp;
6ddc0822	135	struct netns_frags *nf;
d433673e	136	unsigned int sum, sum_truesize = 0;
1e4b8287	137
547b792c IJ	138	WARN_ON(!(q->last_in & INET_FRAG_COMPLETE));
547b792c IJ	139	WARN_ON(del_timer(&q->timer) != 0);
1e4b8287 PE	140
	141	/* Release all fragment data. */
	142	fp = q->fragments;
6ddc0822	143	nf = q->net;
1e4b8287 PE	144	while (fp) {
	145	struct sk_buff *xp = fp->next;
	146
d433673e JDB	147	sum_truesize += fp->truesize;
d433673e JDB	148	frag_kfree_skb(nf, f, fp);
1e4b8287 PE	149	fp = xp;
1e4b8287 PE	150	}
d433673e	151	sum = sum_truesize + f->qsize;
1e4b8287	152	if (work)
d433673e JDB	153	*work -= sum;
d433673e JDB	154	sub_frag_mem_limit(q, sum);
1e4b8287	155
c9547709 PE	156	if (f->destructor)
	157	f->destructor(q);
	158	kfree(q);
1e4b8287 PE	159
	160	}
	161	EXPORT_SYMBOL(inet_frag_destroy);
8e7999c4	162
6b102865	163	int inet_frag_evictor(struct netns_frags nf, struct inet_frags f, bool force)
8e7999c4 PE	164	{
	165	struct inet_frag_queue *q;
	166	int work, evicted = 0;
	167
6b102865	168	if (!force) {
d433673e	169	if (frag_mem_limit(nf) <= nf->high_thresh)
6b102865 AW	170	return 0;
	171	}
	172
d433673e	173	work = frag_mem_limit(nf) - nf->low_thresh;
8e7999c4	174	while (work > 0) {
3ef0eb0d JDB	175	spin_lock(&nf->lru_lock);
3ef0eb0d JDB	176
3140c25c	177	if (list_empty(&nf->lru_list)) {
3ef0eb0d	178	spin_unlock(&nf->lru_lock);
8e7999c4 PE	179	break;
	180	}
	181
3140c25c	182	q = list_first_entry(&nf->lru_list,
8e7999c4 PE	183	struct inet_frag_queue, lru_list);
8e7999c4 PE	184	atomic_inc(&q->refcnt);
3ef0eb0d	185	spin_unlock(&nf->lru_lock);
8e7999c4 PE	186
8e7999c4 PE	187	spin_lock(&q->lock);
bc578a54	188	if (!(q->last_in & INET_FRAG_COMPLETE))
8e7999c4 PE	189	inet_frag_kill(q, f);
	190	spin_unlock(&q->lock);
	191
	192	if (atomic_dec_and_test(&q->refcnt))
	193	inet_frag_destroy(q, f, &work);
	194	evicted++;
	195	}
	196
	197	return evicted;
	198	}
	199	EXPORT_SYMBOL(inet_frag_evictor);
2588fe1d	200
ac18e750 PE	201	static struct inet_frag_queue inet_frag_intern(struct netns_frags nf,
ac18e750 PE	202	struct inet_frag_queue qp_in, struct inet_frags f,
9a375803	203	void *arg)
2588fe1d PE	204	{
	205	struct inet_frag_queue *qp;
	206	#ifdef CONFIG_SMP
2588fe1d	207	#endif
9a375803	208	unsigned int hash;
2588fe1d PE	209
2588fe1d PE	210	write_lock(&f->lock);
9a375803 PE	211	/*
	212	* While we stayed w/o the lock other CPU could update
	213	* the rnd seed, so we need to re-calculate the hash
	214	* chain. Fortunatelly the qp_in can be used to get one.
	215	*/
	216	hash = f->hashfn(qp_in);
2588fe1d PE	217	#ifdef CONFIG_SMP
	218	/* With SMP race we have to recheck hash table, because
	219	* such entry could be created on other cpu, while we
	220	* promoted read lock to write lock.
	221	*/
b67bfe0d	222	hlist_for_each_entry(qp, &f->hash[hash], list) {
ac18e750	223	if (qp->net == nf && f->match(qp, arg)) {
2588fe1d PE	224	atomic_inc(&qp->refcnt);
2588fe1d PE	225	write_unlock(&f->lock);
bc578a54	226	qp_in->last_in \|= INET_FRAG_COMPLETE;
2588fe1d PE	227	inet_frag_put(qp_in, f);
	228	return qp;
	229	}
	230	}
	231	#endif
	232	qp = qp_in;
b2fd5321	233	if (!mod_timer(&qp->timer, jiffies + nf->timeout))
2588fe1d PE	234	atomic_inc(&qp->refcnt);
	235
	236	atomic_inc(&qp->refcnt);
	237	hlist_add_head(&qp->list, &f->hash[hash]);
e5a2bb84	238	nf->nqueues++;
2588fe1d	239	write_unlock(&f->lock);
3ef0eb0d	240	inet_frag_lru_add(nf, qp);
2588fe1d PE	241	return qp;
2588fe1d PE	242	}
e521db9d	243
ac18e750 PE	244	static struct inet_frag_queue inet_frag_alloc(struct netns_frags nf,
ac18e750 PE	245	struct inet_frags f, void arg)
e521db9d PE	246	{
	247	struct inet_frag_queue *q;
	248
	249	q = kzalloc(f->qsize, GFP_ATOMIC);
	250	if (q == NULL)
	251	return NULL;
	252
54db0cc2	253	q->net = nf;
c6fda282	254	f->constructor(q, arg);
d433673e JDB	255	add_frag_mem_limit(q, f->qsize);
d433673e JDB	256
e521db9d PE	257	setup_timer(&q->timer, f->frag_expire, (unsigned long)q);
	258	spin_lock_init(&q->lock);
	259	atomic_set(&q->refcnt, 1);
	260
	261	return q;
	262	}
c6fda282	263
ac18e750	264	static struct inet_frag_queue inet_frag_create(struct netns_frags nf,
9a375803	265	struct inet_frags f, void arg)
c6fda282 PE	266	{
	267	struct inet_frag_queue *q;
	268
ac18e750	269	q = inet_frag_alloc(nf, f, arg);
c6fda282 PE	270	if (q == NULL)
	271	return NULL;
	272
9a375803	273	return inet_frag_intern(nf, q, f, arg);
c6fda282	274	}
abd6523d	275
ac18e750 PE	276	struct inet_frag_queue inet_frag_find(struct netns_frags nf,
ac18e750 PE	277	struct inet_frags f, void key, unsigned int hash)
56bca31f	278	__releases(&f->lock)
abd6523d PE	279	{
abd6523d PE	280	struct inet_frag_queue *q;
5a3da1fe	281	int depth = 0;
abd6523d	282
b67bfe0d	283	hlist_for_each_entry(q, &f->hash[hash], list) {
ac18e750	284	if (q->net == nf && f->match(q, key)) {
abd6523d PE	285	atomic_inc(&q->refcnt);
	286	read_unlock(&f->lock);
	287	return q;
	288	}
5a3da1fe	289	depth++;
abd6523d PE	290	}
	291	read_unlock(&f->lock);
	292
5a3da1fe HFS	293	if (depth <= INETFRAGS_MAXDEPTH)
	294	return inet_frag_create(nf, f, key);
	295	else
	296	return ERR_PTR(-ENOBUFS);
abd6523d PE	297	}
abd6523d PE	298	EXPORT_SYMBOL(inet_frag_find);
5a3da1fe HFS	299
	300	void inet_frag_maybe_warn_overflow(struct inet_frag_queue *q,
	301	const char *prefix)
	302	{
	303	static const char msg[] = "inet_frag_find: Fragment hash bucket"
	304	" list length grew over limit " __stringify(INETFRAGS_MAXDEPTH)
	305	". Dropping fragment.\n";
	306
	307	if (PTR_ERR(q) == -ENOBUFS)
	308	LIMIT_NETDEBUG(KERN_WARNING "%s%s", prefix, msg);
	309	}
	310	EXPORT_SYMBOL(inet_frag_maybe_warn_overflow);